2 TRACK CLEANUP AND TECHNOLOGY TRANSFER Waste to Energy
BIOENERGY PRODUCTION FROM CODIGESTION OF FOOD WASTE AND DAIRY MANURE
1 1
H.M. El-Mashad, 2J.A. McGarvey, and 1R. Zhang
Biological and Agricultural Engineering Department, University of California, One Shields Avenue Davis, CA 95616, USA ; Phone: [phone]; Fax: [phone]; E-mail: [e-mail]. 2U.S. Department of Agriculture, Western Regional Research Center, Agricultural Research Service, Foodborne Contaminants Research Unit, Albany, CA; Phone: [phone]; E-mail: [e-mail]
Abstract The performance of continuously mixed anaerobic digesters were experimentally evaluated for treating manure, food waste, and their mixtures at 372oC and a hydraulic retention time of 20 days. The first mixture was composed of 32% and 68%, and the second was composed of 48% and 52%, of food waste and dairy manure, respectively. The percentage was based on volatile solids (VS). Digesters treating manure and the two mixtures showed stable performance at an organic loading rate (OLR) of 4 gVS/L. However, the digester treating food waste was not stable even at an OLR of 2 gVS/L, as indicated by high volatile fatty acid concentrations and low pH. The pH adjustment by adding hydrate lime to the food waste increased biogas production rate and yield. The 16S rDNA analysis results showed that the food-waste-derived library contained statistically greater numbers of clones related to the phyla Thermotogae and Actinobacteria; and the manure-derived library contained greater amounts of clones related to the phyla Firmicutes, Bacteriodetes, and Spirochetes. The archaeal population structure differed little between digester feed types and was composed of hydrogenotrophic, acetotrophyic, and methylotrophic methanogens. Results of this study showed that adding food waste into a dairy digester could significantly increase the energy production potential and improve the economics of the digester system. Key words: dairy manure, food waste, biogas, renewable energy, microbial analysis Introduction Animal manures and municipal organic wastes are two resources for bioenergy production using anaerobic digestion technologies. Besides energy production, anaerobic digestion reduces the negative environmental impact resulting from direct land application and/or landfilling of untreated waste streams. However, many large dairy manure digesters do not have attractive economics with respect to investment returns due to the low biodegradability of dairy manure. Co-digestion of dairy manure and food waste is considered to be one of the effective options to improve the economics of dairy digesters, because food waste is a highly biodegradable substrate as compared with manure (Zhang et al., 2006). Co-digestion refers to the digestion of two or more substrates in the same digester. It can offer many positive economical and technological benefits (Mata-Alvarez et al., 2000;
186
TRACK 2 CLEANUP AND TECHNOLOGY TRANSFER Waste to Energy
Mshandete et al., 2004; Parawira et al., 2004). Angelidaki and Ellegaard (2003) mentioned that an easily biodegradable substrate, such as food waste, would increase the concentration of active biomass in the digester, which in turn might have a higher tolerance to inhibitory compounds. Moreover, the presence of some inorganic complex (e.g. clay and iron compounds) in some substrates may overcome the inhibitory effects of ammonia and sulfide. Co-digestion of animal manure with other wastes has been studied by other researchers. According to Angelidaki and Ellegaard (2003), animal manure is an excellent carrier substrate when it is co-digested with some concentrated industrial wastes. Some suggested reasons for the suitability of manure as a carrier feedstock are as follows: 1. Manure has a relatively high moisture content (95-88% w.b.) that can help in the pumping and mixing processes. 2. It has a high buffering capacity that helps keep pH in the desired range for methanogenic archaea. 3. It contains all nutrients required for bacterial growth, which is essential for the digestion of industrial waste that has limited nutrient contents. El-Mashad and Zhang (2006a) studied mesophilic batch co-digestion of dairy manure and food waste and determined biogas yields from manure, food waste, and two mixtures. The first mixture was composed of 32% food waste and 68% dairy manure. The second mixture was composed of 48% food waste and 52% dairy manure. The percentage was based on volatile solids (VS). Biogas yields after 30 days of digestion were determined to be 366, 657, 455, and 505 L/kgVS for manure andfood waste, the first and second mixtures with average methane content of the biogas being 66%, 54%, 62%, and 59%, respectively. In another study, El-Mashad and Zhang (2006b) evaluated the performance of completely mixed digesters, treating the two mixtures at 37oC. The digesters fed with the two mixtures were stable at organic loading rates (OLRs) of 2 and 4 gVS/L.day. In that study, microbial structural analyses were not performed, and the experimental biogas yields of the mixtures were compared with the calculated yield of the manure using the model developed by Chen and Hashimoto (1978). This study was a follow-up study with the following objectives: (1) to compare the performance of completely mixed digesters for treating dairy manure and food waste individually and for treating the two mixtures of both substrates, (2) to calculate the
187
TRACK 2 CLEANUP AND TECHNOLOGY TRANSFER Waste to Energy
increase in electricity production potential from a dairy digester after adding food waste, and (3) to characterize the microbial population structure in the digesters that treat the different substrates. Materials and Methods In order to compare the performance of digesters that treat the manure or food waste individually or their mixtures, two mesophilic (372oC) completely mixed digesters, operated at 20-day hydraulic retention time (HRT), were tested in this study on the digestion of dairy manure and food waste. Detailed descriptions of the experimental setup for the digesters can be found in the paper by El-Mashad and Zhang (2006b). One digester was fed with dairy manure and the other was fed with food waste, each at an organic loading rate (OLR) of 4 gVS/L.day. After 20 days of operation, the second digester showed reductions in both biogas production and pH of its effluent and consequently the OLR was decreased to 2 gVS/L.day. When these two digesters were started, they were seeded with the effluent of two lab-scale digesters treating dairy manure. The digester headspaces were flushed with helium gas for five minutes to create anaerobic conditions. The digesters were allowed to sit for two days before feeding began. Test results from the two digesters treating dairy manure or food waste were compared with results of the digesters treating the mixtures of these two substrates (ElMashad and Zhang, 2006b). The food waste and dairy manure were the same for both studies. Substrates characteristics Dairy manure and food wastes used in this study were collected and characterized as described previously (Zhang et al., 2006; El-Mashad and Zhang, 2006b). The total solids (TS) and volatile solids (VS) contents (g/kg) of raw feedstock are shown in Table 1. The daily feeding volume was determined by adding the required amount of VS to give the desired OLR. Measurements and chemical analysis Daily biogas production from each digester was measured using a wet-tip meter. After each digester reached steady state, a biogas sample and an effluent sample were taken from the digester on three consecutive days. The biogas samples were analyzed for CH4 and
188
TRACK 2 CLEANUP AND TECHNOLOGY TRANSFER Waste to Energy
CO2 contents using a gas chromatograph (HP 5890 A) equipped with a thermal conductivity detector as described by Zhang et al. (2006). The digester effluent samples were analyzed for TS, VS, and total dissolved solids (TDS) using standard methods (APHA, 1998). The TDS, soluble chemical oxygen demand (CODs), and volatile fatty acids (VFA) measurements were determined using the methods described by El-Mashad and Zhang (2006b). The pH and electrical conductivity (EC) of the effluents were measured with an Accumet pH/EC meter. Microbial population structure analysis In order to determine the bacterial and archaeal population structure within the digesters, 16S rDNA libraries were constructed, sequenced, and BLAST analyzed as described previously (McGarvey et al., 2005). Briefly, once the digesters reached steadystate conditions, samples were taken on three different days and DNA was extracted from them. The 16S rDNA sequences contained within these extracts were amplified via the polymerase chain reaction (PCR) using the eubacterial-specific primers 27f (5 AGAGTTTGATCCTGGCTCAG 3) and 1392r (5 GACGGGCGGTGTGTAC 3) as described by Lane (1991), or the archaea specific-primers 25f (5'CYGGTTGATCCTGCCRG-3') and 1492r (5'-GGTTACCTTGTTACGACTT-3') as described by Dojka et al. (1998 and 2000). The PCR products were purified by ethanol precipitation, cloned into plasmid vectors using the Qiagen PCR Cloning Kit (Qiagen, Valencia CA) as per the manufacturers instructions, and transformed into E. coli TOP10F cells (Invitrogen, Carlsbad, CA) by heat shock (42oC for 30 sec). Clones were plated on LB agar plates containing kanamycin (Km) (50 g ml-1), isopropyl- -D thiogalactopyranoside g (IPTG) (20 mM), and 5-bromo-4-chloro-3-indolyl- -D-galactopyranoside (X-gal) (80
ml-1). White colonies were selected and grown in 96 well plates in LB Km broth. Plasmids containing the 16s rDNA sequences were amplified using the TempliPhi 100 Amplification Kit (Amersham Biosciences, Sunnyvale, CA) as per the manufacturers instructions. Sequencing reactions were performed using the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA), using the primer 1392r for bacteria or 1942r for archaea. Sequencing reactions were purified using the DyeEx 96 Kit (Qiagen, Valencia, CA); electrophoresis and readout were performed using an AppliedBiosystems 3100 Genetic Analyzer (Applied Biosystems, Foster City, CA). DNA sequences were edited manually to
189
TRACK 2 CLEANUP AND TECHNOLOGY TRANSFER Waste to Energy
correct falsely called bases and trimmed at both the 5 and 3 ends using Chromas software (Technelysium Pty. Ltd., Helensvale Australia). Only sequences with unambiguous reads longer than 500 bp were used, and each read averaged approximately 600 bp. The predicted 16S rDNA sequences from this study were compared to 16S rDNA sequences in a BLASTable database constructed previously (McGarvey et al., 2004). This database contains sequences downloaded from the Ribosomal Database Project II (http://rdp.cme.msu.edu/download/SSUrRNA/unaligned/); Release 8.1. Comparisons were made using the program BLASTALL (ftp://ftp.ncbi/nih.gov/toolbox/ncbi_tools/) and a FASTA-formatted file containing the predicted 16S rDNA sequences. Operational taxonomic units (OTUs) were defined as clones with > 97% sequence identity. Comparisons of the 16S rDNA libraries were analyzed using Library Compare software (available at http://rdp.cme.msu.edu/comparison/comp.jsp), which estimates the likelihood that the frequency of membership in a given taxon is the same for the two libraries using the equation:
where N1 and N2 are the total number of sequences for library 1 and 2, respectively, and x and y are the number of sequences assigned to an OTU from library 1 and 2, respectively. Results and Discussion Performance of the digesters Average daily biogas and methane production rates and yields of different digesters are shown in Table 2. A steady-state operation was assumed after 60 days (3 HRTs) for the digesters. As can be seen, biogas production rates were 1.26, 0.51, 1.91, and 2.02 L/L.day for the digesters treating manure, food waste, and the first and second mixtures, while the biogas yields were calculated to be 314, 256, 476, and 504 L/kgVS, respectively. Average methane content of the biogas produced was 58%, 45.7%, 58.5% and 63.3%, respectively. The digesters treating the manure and the two mixtures were stable as determined by low VFA concentrations and relatively neutral pH values (Table 2). The total VFA concentration in the effluent of the manure digester was 678 mg [acetic acid]/L, which is
190
TRACK 2 CLEANUP AND TECHNOLOGY TRANSFER Waste to Energy
comparable to the VFA concentrations observed in other dairy-waste digester effluents (Martin, 2003). Acetic acid represented approximately 46% of the total VFA content. The digester treating food waste was unstable at an OLR of 4 g VS/L.day (data not shown), thus the OLR was reduced to 2 g VS/L. day. Although the OLR was reduced, the digester continued to suffer from high VFA concentrations (6778 mg [acetic acid]/L), which resulted in low pH values (5.86) after two months of operation. Acetic acid was the major constituent of the VFAs (about 45%). Compared to the other digesters, the food-waste digester had a higher concentration of CODdis. The accumulation of these intermediates (i.e. VFA and other organic compounds) indicated that the hydrolytic and acidogenic bacteria had a higher tolerance to low pH values and high VFA concentrations than the methanogenic archaea did (Cho et al., 1995). After characterizing effluent the from the food-waste digester, the pH was increased to about 7.1 by adding hydrated lime, Ca (OH)2. This addition resulted in increased biogas production. After about one month, the average biogas production rate and yield were 1.29 L/L day and 643.2 L/kg VS. The average methane content of the biogas was measured to be 64%. Although adjusting the pH with hydrated lime was effective in maintaining the pH within the proper range for methanogenic archaea, the amount of the lime added should be controlled in order to prevent the inhibitory effects of high calcium concentrations to methanogenic archaea (McCarty, 1964). The measured VS reductions in the digesters were calculated to be 36.1%, 57.%, 42.2%, and 57.1% for manure, food waste, and first and second mixtures, respectively. The VS reduction in the manure digester is typical for dairy digesters (Dugba and Zhang, 1999). The measured, total dissolved solids were 15.8, 5.27, 17.7, and 15.3 g/L and the EC values were 5.78, 5.05, 5.7 and 6.53 mS/cm for the effluent of the digesters, respectively. It should be mentioned, that although the TDS and EC are commonly used as parameters for determining the salt contents of water, they do not accurately measure the salt concentrations in digester effluents. This is because the TDS in the digester effluents contain organic as well as inorganic matter. A more detailed study is required to determine the relationships between the TDS and the salt concentrations in the digester effluents.
191
TRACK 2 CLEANUP AND TECHNOLOGY TRANSFER Waste to Energy
Microbial population structure analysis Bacterial 16S rDNA libraries containing approximately 500 clones were constructed from total DNA isolated from the effluents of the anaerobic digesters treating food waste, manure, and the two mixtures. Analysis of these data revealed significant differences between the libraries derived from the digesters treating food waste and manure (Table 3). Specifically, it was observed that the food-waste-derived library contained statistically greater numbers of clones related to the phyla Thermotogae and Actinobacteria, and the manure-derived library contained more clones related to the phyla Firmicutes, Bacteriodetes, and Spirochetes. Co-digestion of food waste and manure in the two mixtures resulted in few differences between the resultant libraries, except for the greater abundance of sequences related to the Chloroflexi in the second digester and the small but statistically significant greater abundance of the Bacteriodetes in the digester treating the first mixture. However, when compared to the libraries derived from the effluents of the food waste or manuredigester effluents, they resembled a conglomeration of the two, in that both mixed-digestereffluent-derived libraries contain sequences related to the phyla Thermotogae and Firmicutes at levels that are intermediate to the food waste and manure-digester-effluent-derived libraries. Interestingly, both mixed-food waste and manure-digester-effluent-derived libraries showed significantly greater levels of the phylum Chloroflexi than either manure or foodwaste-effluent derived libraries. However, it is unclear what specific selection parameter influenced their increase. Archaeal 16S rDNA libraries were also constructed for each digester type. Analysis of these libraries revealed much lower levels of diversity than that observed in the bacterial libraries, with only one phylum, the Eurychaeota, present in each library (Table 4). Comparison of the libraries derived from the food-waste-digester effluent and the manuredigester effluent revealed small but significant differences between the orders Methanosarcinales and Methanomicrobiales. However, when the mixtures were digested, no significant differences were observed between the two digester effluents at the order level. It should also be noted that in all of the libraries, the most numerous operation taxonomic units identified were related to the order Methanosarcinales, followed by the Methanomicrobiales
192
TRACK 2 CLEANUP AND TECHNOLOGY TRANSFER Waste to Energy
and the Methanobacteriales. Thus, the archaeal population structure differed little between digester feed type and was composed of a mixture of hydrogenotrophic, acetotrophyic, and methylotrophic methanogens. Energy potential and economic analysis Energy production potential from a dairy digester treating manure from 1,800 cows with and without adding food waste is shown in Fig.1. Without adding food waste, a daily electricity production of 4,667 kWh is predicted, which is equivalent to about 2.6 kWh/cow. day. After the addition of 19.1 or 38.2 tons (wet basis) of food waste a day, the daily electricity production potential would increase to 10,767 or 16,465 kWh/day, respectively, which represents the increase in the electricity production potential by approximately 2.3 or 3.5-fold , compared to the electricity potential for digesting the manure alone. It should be mentioned that in the pervious study by El-Mashad and Zhang (2006b), the electricity production potential from the same amount of manure was calculated to be 5313 kWh/day, based on the methane yield calculated from the model of Chen and Hashimoto (1978), which is higher in the yield determined from this study. A preliminary economic analysis showed that based the tipping fee of food waste ($25 per ton (wet)) and biogas production of food waste, the payback period of the dairy digesters could be reduced from nine years for digesting manure alone to four years for codigesting food waste and manure. A more detailed design of the digester for co-digestion of food waste and manure and its economic analysis are under investigation. Conclusions Based on the results obtained in this study, the following conclusions can be drawn: 1. Anaerobic digesters treating mixtures of dairy manure and food waste were stable at organic loading rates of 4 g VS/L.day, while the digester treating food waste alone was not stable even at 2 gVS/L.day. High VFA concentrations and low pH were found in the food-waste digesters as a result of the high biodegradation rate of food waste. 2. The microbial population structure of food-waste-digester effluent contained statistically greater numbers of clones related to the phyla Thermotogae and Actinobacteria, and the manure-digester effluent contained greater amounts of clones
193
TRACK 2 CLEANUP AND TECHNOLOGY TRANSFER Waste to Energy
related to the phyla Firmicutes, Bacteriodetes, and Spirochetes. The archaeal population structure differed little between digester feed types and was composed of hydrogenotrophic, acetotrophyic, and methylotrophic methanogens. 3. Adding food waste into the dairy manure-digester would significantly increase energy production potential and improve the economics of the digester system. Acknowledgment This research was supported with a grant from Sacramento Municipal Utility District (SMUD), and the USDA, ARS, FCR. References Angelidaki, I., and L. Ellegaard, 2003. Codigestion of Manure and Organic Wastes inCentralized Biogas Plants, Applied Biochemistry and Biotechnology, 109, pp. 95105. Chen, Y.R., and A.G. Hashimoto, 1978. Kinetics of Methane Fermentation.Biotechnology and Bioegineering Symposium, No. 8, pp. 269-82. Cho, J.K., S.C. Park, and H.N. Chang, 1995. Biochemical Methane Potential and Solid-State Anaerobic Digestion of Korean Food Wastes, Bioresource Technology, 52,pp. 245-253. Dojka, M.A., P. Hugenholtz, S.K.Haack, and N.R. Pace, 1998. Microbial Diversity in a Hydrocarbon and Clorinatedsolvent-Contaminated Aquifer Undergoing Intrinsic Bioremediation, Appl. Environ. Microbiol. 64, pp. 38693877. Dojka, M.A., J.K. Harris, and N.R. Pace, 2000. Expanding the Known Diversity and Environmental Distribution of Uncultured Phylogenetic Division of Bacteria, Appl. Environ. Microbiol 66, pp. 16171621. Dugba, P.N., and R. Zhang, 1999. Treatment of Dairy Wastewater with Two-Stage Anaerobic Sequencing Batch Reactor Systems-Thermophilic Versus Mesophilic Oper...